Re: Questions about indexes?

Ryan Bradetich <rbradetich(at)uswest(dot)net> writes:
> Although the table schema is immaterial, I will provide it so we have a
> common framework for this discussion:

> host_id integer (not null)
> timestamp datetime (not null)
> category text (not null) [<= 5 chars]
> anomaly text (not null) [<= 1024 chars]

> This table is used to store archived data, so each row in the table must
> be unique. Currently I am using a primary key across each column to
> enforce this uniqueness.

It's not real clear to me why you bother enforcing a constraint that the
complete row be unique. Wouldn't a useful constraint be that the first
three columns be unique? Even if that's not correct, what's wrong with
tolerating a few duplicates? You can't tell me it's to save on storage
;-)

> I am not sure why all the data is duplicated in the index ... but i bet
> it has to do with performance since it would save a lookup in the main
> table.

An index that can't prevent looking into the main table wouldn't be
worth anything AFAICS ...

regards, tom lane